I wasn’t trying to reduce screen time but trying to drink more water.

So I built an iPhone app (SwiftUI) that locks selected apps using Apple’s Screen Time APIs. To unlock them, ThirstTrapp activates the front camera and watches for ~15 seconds while I drink.

On-device vision checks for:

   - a face
   - a drinking container (glass/bottle)
   - and a plausible drinking gesture (container overlapping mouth region over time)

If that sequence is detected for ~15 seconds, the apps unlock for a configurable window (e.g. 2 hours). Then they lock again until the next drink.

It doesn’t try to stop me from scrolling. It just makes hydration the entry fee. And since I'm addicted to scrolling I will actually drink. It’s slightly absurd, but it’s been more effective for me than reminder apps ever were.

All processing happens on-device. No images or video leave the phone or are stored. No account registration or login.

Curious what you think - especially from a technical perspective.